Our Content Editor, Gabriel Barnes, speaks to George Zarkadakis, Digital Lead at Willis Towers Watson and author of Cyber Republic: reinventing democracy in the age of intelligent machines. They discuss how Artificial Intelligence is influencing politics and George gives his predictions for the future.
Gabriel Barnes: You talk a lot about the influence that Artificial Intelligence has on politics. How is this happening, and could it be perceived as a threat to democracy?
George Zarkadakis –There are multiple areas where AI is impacting our politics and the trust that citizens bestow on democratic political institutions. The first area is in how information and content are filtered by AI algorithms in social media. There, because of the dominant advertising model of social media platforms, the result is the propagation of polarized views and fake news, and the creation of “echo chambers” where biases are confirmed – and never challenged.
The second area is in the automation of human judgement in various government processes, for example in the allocation of welfare benefits, healthcare provision, approval of parole, shortlisting suspects in police databases, etc.; as AI algorithms are trained in historical data that include social biases, they tend to propagate these biases and amplify social injustice.
The third area is surveillance and how personal data are used by government agencies. Unless the use and governance of AI become transparent and participatory the impact of AI systems in all those three areas will be very detrimental to the future of democracy.
Gabriel Barnes: When we look to the future of Artificial Intelligence and politics, is it all bad news?
George Zarkadakis – AI is a cognitive multiplier that can enable citizens to better understand the issues, organize, participate, and influence the political process. Unfortunately, there is currently scant investment and innovation in how we can use AI to improve our politics.
Gabriel Barnes: How do you predict Artificial Intelligence might advance in the near future?
George Zarkadakis – I predict that AI will embed in almost every system that operates using software. Advances in hardware chips that specialize in AI processing suggest that the future is the “intelligent internet of things”. Further advances in transfer learning will deliver AI systems that will behave as if they have “general intelligence”, i.e., they will be able to independently seek, collect, and process knowledge in order to solve hard problems.
It will be a very exciting future for science, as we recently saw in how AlphaFold, an AI system developed by Deep Mind, found a way to predict how proteins fold, a 50-year-old grand challenge in biology.
Gabriel Barnes: How do you think these advances might be received? Should they be feared or encouraged?
George Zarkadakis – Fear usually results from ignorance, and I believe this is the case for AI. Most people fear AI systems because either they are not a part of how those systems are evolving, or they feel like they don’t have a say in how those systems are deployed in society. People are worried about their jobs, their future, and their rights in a world where AI belongs to powerful minorities and a handful of omnipotent corporations. For AI to deliver equitable benefits to society we need to find ways for citizens to participate in the governance of data and algorithms.
In my book “Cyber Republic: reinventing democracy in the age of intelligent machines (MIT Press)”, I discuss several proposals to democratize the AI economy, such as Data Trusts and Data Cooperatives, as well as participatory business models for digital platforms using cryptoeconomics and cryptocurrencies.
